Most patients with inflammatory bowel disease (IBD), comprising

Transcription

1 FUTURE DIRECTIONS AND METHODS IN IBD RESEARCH Quality of Care and Outcomes Among Hospitalized Inflammatory Bowel Disease Patients: A Multicenter Retrospective Study Geoffrey C. Nguyen, MD, PhD,*, Sanjay K. Murthy, MD, MSc,, Brian Bressler, MD, MSc, Mindy C. W. Lam, MD, Ali Alali, MD,* Asmae Toumi, BSc, k Jason Reinglas, MD, Adam Rampersad, BSc, k Adam V. Weizman, MD, MSc,* and Waqqas Afif, MD, MSc, k on behalf of the CINERGI group Background: Half of patients with inflammatory bowel disease (IBD) require hospitalization. We sought to characterize inpatient quality indicators of care and outcomes during IBD-related hospitalizations at 4 major IBD referral centers in Canada. Methods: We conducted a multicenter retrospective cohort study of patients with IBD admitted from 2011 to 2013 to tertiary centers in Toronto, Montreal, Ottawa, and Vancouver. We assessed the following inpatient indicators of care: pharmacological venous thromboembolism (VTE) prophylaxis, Clostridium difficile testing, and medical rescue therapy for steroid-refractory ulcerative colitis (UC). We also evaluated rates of VTE, C. difficile infection, and IBD-related surgery. Results: There were 837 patients hospitalized for IBD (Crohn s disease, 59%; UC, 41%). The proportion of patients with IBD who received VTE prophylaxis and C. difficile testing were 77% and 82%, respectively, although these indicators varied significantly by center and admitting specialty. Patients admitted under surgeons were more likely than those admitted under gastroenterologists to receive VTE prophylaxis (84% versus 74%, P ¼ 0.016) but less likely to be tested for C. difficile (41% versus 88%, P, ). The rate of VTE was the same for those who did and did not receive VTE prophylaxis (2.2 per 1000 hospital-days). Among the 14 VTE events, 79% had received prophylaxis, but only 36% within 24 hours of admission. Among steroid-refractory UC patients, 70% received rescue therapy within 7 days of steroid initiation. The proportion of patients with UC and CD who required respective bowel surgery was 18% and 20%, respectively. Conclusions: There are opportunities to optimize quality of care among hospitalized patients with IBD. (Inflamm Bowel Dis 2017;23: ) Key Words: venous thromboembolism, Crohn s disease, ulcerative colitis, Clostridium difficile, surgery Most patients with inflammatory bowel disease (IBD), comprising Crohn s disease (CD) and ulcerative colitis (UC), are managed in the outpatient setting. However, hospitalization is required in half of patients during their disease course. 1 Reports suggest that hospitalizations for IBD are rising in North America and occur more often early in the course of the disease. 2 Within the first 2 years of diagnosis, 24% of patients with CD and 20% of patients with UC require an inpatient stay. 3 During hospitalization, patients with IBD are at increased risk of venous thromboembolic events (VTE) and Clostridium Received for publication November 29, 2016; Accepted January 29, From the *Mount Sinai Hospital Centre for Inflammatory Bowel Disease, Division of Gastroenterology, University of Toronto, Toronto, Ontario, Canada; Institute for Health Policy Management and Evaluation, University of Toronto, Toronto, Ontario, Canada; Department of Medicine, The Ottawa Hospital IBD Centre, University of Ottawa, Ottawa, Ontario, Canada; Division of Gastroenterology, St. Paul s Hospital, University of British Columbia, Vancouver, British Columbia, Canada; and k Division of Gastroenterology, McGill University, Montreal, Québec, Canada. S. K. Murthy: Abbvie, Takeda, Shire, Ferring (Advisor/Speaker). B. Bresslerv: Shire, Ferring, Janssen, Abbvie, Takeda, Actavis (Advisor/Speaker); Amgen, Pendopharm, Genentech, Celltrion, Pfizer, Allergan (Advisor); Janssen, Abbvie, GSK, BMS, Amgen, Genentech, Merck, RedHill Biopharm, BI, Qu Biologic, Celgene, Alvine (Research Support); Qu Biologic (Stock options). A. V. Weizman: Janssen, Abbvie, Shire, Takeda (Advisor/Speaker). W. Afif: Abbvie, Janssen, Takeda, Shire, Ferring, Merck, Pfizer (Advisor); Abbvie, Janssen, Prometheus, Theradiag (Research Support). G. C. Nguyen, B. Bresslerv, S. K. Murthy, A. V. Weizman, and W. Afif are practicing gastroenterologists at their respective institutions involved in this study. The remaining authors have no conflict of interest to disclose. The Canadian IBD NEtwork for Research and Growth in Quality Improvement (CINERGI) group includes the following: G. C. Nguyen (University of Toronto); J. L. Jones (Dalhousie University); W. Afif (McGill University); B. Bressler (University of British Columbia); S. Fowler (University of Saskatchewan); S. Halder (McMaster University); V. W. Huang (University of Alberta); R. Khanna (Western University); S. K. Murthy (University of Ottawa); Joannie Ruel (University of Sherbrooke); C. H. Seow (University of Calgary); L. E. Targownik (University of Manitoba); A. V. Weizman (University of Toronto). Address correspondence to: Geoffrey C. Nguyen, MD, PhD, 600 University Avenue, Suite 437, Toronto, ON M5G 1X5, Canada ( geoff.nguyen@utoronto.ca). Copyright 2017 Crohn s & Colitis Foundation DOI /MIB Published online 21 March Inflamm Bowel Dis Volume 23, Number 5, May

2 Nguyen et al Inflamm Bowel Dis Volume 23, Number 5, May 2017 difficile infection, which are associated with greater than 2-fold and nearly 4-fold increase in-hospital mortality, respectively. 4 6 Accordingly, clinical practice guidelines recommend pharmacological VTE prophylaxis and testing for C. difficile for patients hospitalized for IBD. 7,8 Moreover, the American Gastroenterological Association has included VTE prophylaxis and C. difficile testing as inpatient quality indicators in its IBD Performance Measure Set. 9 Furthermore, nearly two-thirds of patients who are hospitalized for severe UC are unresponsive to intravenous corticosteroids and require rescue therapy with infliximab, cyclosporine, or colectomy. 10 Current clinical guidelines recommend that rescue therapy be considered for inadequate steroid response after 3 days. 7 The administration of rescue therapy within 7 days of steroid initiation has been suggested as an additional inpatient quality indicator in Canada. 11 However, it is unclear whether a higher standard for rescue therapy within 4 days of steroid initiation should be considered. Previous studies from the United States suggest there may be substantial geographic variation in the inpatient management of IBD. 12 We conducted a multicenter study of IBD hospitalizations among tertiary IBD centers in Canada s 4 largest metropolitan areas (Toronto, Montreal, Vancouver, and Ottawa) to determine the frequency to which inpatient quality indicators are being adhered. Secondary aims were to assess for geographic variations in inpatient quality indicators across the 4 centers and to evaluate rates of adverse clinical outcomes. 13 METHODS Data Sources A retrospective chart review of IBD admissions was independently conducted at 4 high-volume referral centers in each of the following major Canadian cities: Toronto, Ontario (Mount Sinai Hospital); Montreal, Quebec (McGill University Health Centre), Ottawa, Ontario (The Ottawa Hospital); and Vancouver, British Columbia (St. Paul s Hospital). Data were extracted from electronic and paper charts, with access to clinical information, laboratory investigations, endoscopic evaluations, pharmacological records, and discharge summaries. Eligibility Criteria We identified admissions with a most responsible diagnosis for IBD using International Classification of Diseases, Ninth Revision and International Classification of Diseases, Tenth Revision codes for UC (556.x and K51, respectively) and CD (555.x and K50, respectively) from 2011 to The medical charts were reviewed to confirm diagnosis. Only patients with IBD who were admitted for treatment of active disease for more than 24 hours were included. We excluded patients with IBD who were admitted electively for diagnostic or surgical procedures or who were admitted for a non-ibd indication. Data Collection Predictor Variables Patient demographic information and clinical characteristics were collected and included: age at admission, sex, IBD diagnosis (CD, UC, or IBD unclassified), and presence of hemodynamically significant gastrointestinal bleeding on admission. We also recorded admission-related factors, including the primary admitting service (gastroenterology, internal medicine, surgery, or other); and weekend versus weekday admissions. When available, the medical chart, phenotype data using the Montreal classification were also abstracted and included disease extent for UC and disease location and behavior for CD. Outcomes Outcomes were assessed through review of the inpatient record including the discharge summary, physician and nursing notes, diagnostic testing and endoscopy reports, and medical administration record. We assessed the following inpatient quality of care indicators: occurrence and timing of VTE prophylaxis; C. difficile testing (performed by polymerase chain reaction at all sites); and medical rescue therapy for steroid-refractory UC administered within 7 days of steroid initiation. In determining whether or not VTE prophylaxis was administered, we excluded admissions in which patients presented with hemodynamically significant bleeding. Among those who did receive pharmacological prophylaxis, we collected additional data on whether it was administered within 24 hours of admission, type of VTE prophylaxis used, and duration of therapy. For evaluation of the rate of C. difficile testing, we included only admissions in which patients presented with diarrhea. We collected data on the following outcomes using standardized collection forms: VTE events, major bleeding events related to VTE prophylaxis, C. difficile infection detected by polymerase chain reaction assay, and bowel resection during hospital admission. We defined major bleeding as: intracranial, intraspinal, or retroperitoneal bleeding; bleeding into any major organ; or postoperative bleeding that lead to reoperation. We also ascertained whether major bleeding required transfusion of packed red blood cells. Statistical Analysis We performed statistical analysis using Stata 14 SE (StataCorp LP, College Station, TX). Descriptive analysis was conducted to compare demographic and clinical characteristics of patients among the 4 participating centers. The chi-square and Fisher s exact tests were used to compare categorical variables, whereas the Student s t test was used to compare continuous variables. Multivariate logistic regression was used to identify predictors of VTE prophylaxis and C. diff testing while simultaneously adjusting for each of the predictors. A robust variance estimator was used to account for clustering by site, service, and physician provider. The unadjusted proportion of patients who received VTE prophylaxis and those who received it within 24 hours were calculated and stratified by primary service. The crude rates of minor and major bleeding were calculated by dividing the 696

3 Inflamm Bowel Dis Volume 23, Number 5, May 2017 Inpatient IBD Quality of Care total number of events by the number of total person-days on prophylactic anticoagulation. For those who did not receive VTE prophylaxis, the rate was calculated using the total person-days in hospital as the denominator. Ethical Considerations The Research Ethics Board at Mount Sinai Hospital (Toronto, Ontario), McGill University Health Centre (Montreal, Quebec), the Ottawa Hospital (Ottawa, Ontario), and St. Paul s Hospital (Vancouver, British Columbia) approved the study protocol. RESULTS There were a total 837 patients hospitalized for IBD during the study period, of which 493 were for CD (59%), 343 were for UC (41%), and 1 (0.1%) was for IBD unclassified. The number of hospitalizations at each specific site was as follows: Toronto, 243 (29%); Montreal, 151 (18%); Ottawa, 271 (32%); and Vancouver, 172 (21%). The clinical and demographic characteristics of subjects at each hospital site are given in Table 1. Most IBD-related hospitalizations (62%) were under the gastroenterologist service though this varied by site. Pharmacological VTE Prophylaxis Pharmacological VTE prophylaxis was administered to 77% of hospitalized patients with IBD, although this varied by location: ranging from 62% in Montreal to 84% in Ottawa (Table 2). Overall, 62% received anticoagulation within 24 hours of admission but this indicator differed significantly by hospital site, ranging from 26% in Montreal to 76% in Ottawa (Table 2). There was also hospital-based variation in preference of pharmacological agents: whereas low molecular weight heparin was strongly favored in Toronto (97%), Montreal (73%), and Ottawa (87%), unfractionated heparin (46%) and low molecular weight heparin (54%) were preferred in Vancouver. Admissions under general surgery were more likely to receive VTE prophylaxis compared with the gastroenterology service (84% versus 74%, P ¼ 0.016; adjusted odds ratio, 2.07; 95% confidence interval, ). However, the VTE prophylaxis rate within 24 hours of admission did not significantly differ by the admitting service (Fig. 1). Interestingly, IBD admissions on the weekend were less likely to receive VTE prophylaxis within 24 hours of admission only if admitted under gastroenterology (51% versus 65%, P ¼ 0.017; adjusted odds ratio, 0.57; 95% confidence TABLE 1. Demographic and Clinical Characteristics of Hospitalized IBD Patients Stratified by Hospital Center Hospital Center Demographic Variables Overall (N ¼ 837) Toronto (N ¼ 243) Montreal (N ¼ 151) Ottawa (N ¼ 271) Vancouver (N ¼ 172) Global P Age (mean 6 SD) Sex Male 391 (47) 98 (41) 83 (55) 126 (47) 84 (49) Female 443 (53) 143 (59) 68 (45) 144 (53) 88 (51) Primary service,0.001 Gastroenterology 520 (62) 150 (62) 57 (38) 189 (70) 124 (72) General internal medicine 142 (17) 52 (22) 36 (24) 45 (17) 9 (5) General surgery 167 (20) 39 (16) 57 (38) 32 (12) 39 (23) Other #5 #5 #5 #5 #5 Weekend admission 150 (18) 39 (16) 21 (14) 65 (24) 25 (15) Bleeding on admission 406 (51) 134 (56) 48 (32) 161 (60) 63 (44),0.001 Hemodynamically significant 63 (8) #5 36 (24) 7 (3) 15 (10),0.001 IBD diagnosis 0.32 UC 343 (41) 104 (43) 53 (35) 113 (42) 73 (42) CD 493 (59) 139 (57) 97 (64) 158 (58) 99 (58) IBD unclassified #5 #5 #5 #5 #5 Comorbidity,0.001 None 768 (92) 232 (95) 151 (100) 233 (86) 152 (89) 1 comorbidity 51 (6) 6 (3) #5 28 (10) 17 (10) $2 comorbidities 17 (2) #5 #5 10 (4) #5 CD presentation,0.001 Inflammatory 183 (42) 81 (59) 15 (21) 77 (49) 10 (15) Stricturing 160 (37) 30 (22) 35 (49) 62 (40) 33 (49) Penetrating 92 (21) 27 (20) 22 (31) 18 (11) 25 (37) 697

4 Nguyen et al Inflamm Bowel Dis Volume 23, Number 5, May 2017 TABLE 2. Quality Indicators and Outcomes of Hospitalized Patients with IBD Stratified by Hospital Center Hospital Center Overall (N ¼ 837) Toronto (N ¼ 243) Montreal (N ¼ 151) Ottawa (N ¼ 271) Vancouver (N ¼ 172) Global P Quality indicators of care VTE prophylaxis 572 (77) 191 (81) 71 (62) 219 (84) 91 (70),0.001 Within 24 hours of admission 457 (62) 154 (65) 30 (26) 198 (76) 75 (58),0.001 C. difficile testing 451 (82) 173 (91) 53 (61) 166 (83) 59 (81),0.001 Rescue therapy within 7 days of steroid initiation (UC only) 53 (70) 20 (61) #5 (75) 30 (77) N/A 0.32 Rescue therapy within 4 days of steroid initiation (UC only) 11 (14) #5 (12) #5 (50) #5 (13) N/A 0.19 Outcomes In-hospital VTE (per 1000 hospital-days) 2.2/ /1000 a 3.8/ /1000 a 4.9/1000 N/A Major bleeding (per 1000 hospital-days) 6.0/1000 0/1000 a,b 1.9/1000 b 17.7/ /1000 N/A C. difficile positive test, n (%) 50 (11) 20 (12) #5 (8) 20 (12) 6 (10) 0.66 Steroid responsiveness, n (%) 135 (54) 55 (54) 24 (52) 56 (55) N/A 0.93 Bowel surgery (CD only, %) 99 (20) 19 (14) 30 (31) 14 (9) 36 (36),0.001 Colectomy (UC only, %) 60 (18) 17 (17) 21 (40) 7 (6) 15 (21),0.001 N/A, not available. a P, 0.05 compared with Vancouver. b P, 0.01 compared with Ottawa. interval, ) but not general medicine or surgery. The presence of gastrointestinal bleeding (not hemodynamically significant) on admission was associated with higher rates of VTE prophylaxis within 24 hours (67% versus 57%, P ¼ 0.005). Safety and Effectiveness of Pharmacological VTE Prophylaxis The rate of major bleeding was not significantly higher in the group who received VTE prophylaxis compared with those who did not (7.6 versus 6.0 per 1000 hospital-days, P ¼ 0.62). The rate FIGURE 1. Inpatient quality indicators of care by specialty of admitting service. The proportion of admissions that received VTE prophylaxis, VTE prophylaxis within 24 hours of admission, C. difficile testing, and initiation of medical rescue therapy (among steroid-refractory UC admissions) within 7 days of steroid start are shown stratified by admitting service (gastroenterology, general internal medicine, or general surgery). Statistically significant differences (P, 0.05) compared with gastroenterology specialty are indicated with an asterisk (*)

5 Inflamm Bowel Dis Volume 23, Number 5, May 2017 Inpatient IBD Quality of Care of major bleeding requiring blood transfusion also did not differ (3.2 versus 3.0 per 1000 hospital-days, P ¼ 0.96). Among those receiving VTE prophylaxis, the rate of major bleeding did vary by hospital site (Table 2), ranging from 0/1000 in Toronto to 17.7/ 1000 in Ottawa. The overall rate of VTE during hospitalization was 2.2 per 1000 hospital-days and was higher in Montreal (3.8/1000 hospitaldays) and Vancouver (4.9/1000 hospital-days) compared with Toronto (0.9/1000 hospital-days) and Ottawa (1.0/1000 hospital-days), where VTE prophylaxis rates were higher. The anatomical location of the 14 VTE events were 29% lower extremities, 21% pulmonary embolism, 36% upper extremities, and 14% other site. However, VTE rate did not significantly differ between those who did and did not receive VTE prophylaxis (2.2 versus 2.2 per 1000 hospital-days, P ¼ 0.95). The VTE rate was lower among those who received VTE prophylaxis within 24 hours of admission compared with those who never received any prophylaxis (1.2 versus 2.2/1000 hospital-days, P ¼ 0.4). Among those who had VTE events during hospitalization, 79% (n ¼ 11/14) had received VTE prophylaxis. Nearly three-quarter (73%) of patients (n ¼ 8/11) who developed VTE while on VTE prophylaxis did not have any surgical procedures during hospitalization. However, only 36% of those with VTE received VTE prophylaxis within 24 hours of admission. Testing for C. difficile Although the overall rate of C. difficile testing among patients with UC and CD with diarrhea was 82%, there was variation with geographic location, ranging from 61% in Montreal to 91% in Toronto (Fig. 1). Testing was similarly high among admissions to gastroenterology and general medicine (88% and 87%, respectively) but considerably lower in those admitted to general surgery (41%, P, ; adjusted odds ratio, 0.12; 95% confidence interval, ). The prevalence of C. difficile positivity among those tested was 11%, and this did not significantly vary with location (Table 2). Medical Therapy in UC Among hospitalized patients with UC treated with systemic steroids, 58% demonstrated adequate clinical response and therefore required neither change in therapy nor surgery. The overall steroid responsiveness rate was 54% and did not vary by site (Table 2); data were not available for Vancouver. Among those with inadequate response to medical therapy, 76 (79%) received medical rescue therapy, and this did not significantly vary by geographic site. Ninety-nine percent of those who required medical rescue therapy received infliximab. Among individuals receiving rescue therapy, 70% received it within 7 days of starting steroids. This proportion was lower in Toronto (61%) than in Montreal (75%) and Ottawa (77%), although this was not statistically significant (P ¼ 0.3). Only 14% of patients with UC received second-line therapy within 4 days of initiating systemic steroids. In Toronto and Ottawa, the proportions were 12% and 13%, respectively, whereas it was 50% in Montreal (P ¼ 0.19). There was no difference between patients admitted to gastroenterology and general internal medicine with respect to receiving rescue therapy within 7 days (70% versus 78%; P ¼ 0.4) or 4 days (13% versus 17%, P ¼ 0.5). Surgery Among hospitalized patients with UC, 18% underwent colectomy during hospitalization. There was substantial variation in the proportion requiring surgery by geographic location: Ottawa (6%), Toronto (17%), Vancouver (21%), and Montreal (40%). Among steroid-refractory UC patients, the colectomy rate was 25% with following site-specific rates: Ottawa (16%), Toronto (33%), and Montreal (40%); P ¼ Nineteen percent of patients with UC who did not respond to steroids underwent colectomy without attempting rescue therapy. Among the remaining patients who did receive rescue therapy, 92% were able to avoid surgery during hospitalization. The overall rate of bowel resection for hospitalized patients with CD was 20%. As in UC, there was substantial variation by geographic location: Ottawa (9%), Toronto (14%), Montreal (31%), and Vancouver (36%). DISCUSSION In this large multicenter retrospective study, we identified significant variations in inpatient quality indicators and outcomes across centers and admitting services. These findings may reflect opportunities to optimize quality of care. Because of the increased mortality associated with VTE and C. difficile infection, processes of care to prevent or promptly treat these complications are cardinal in improving hospital outcomes. Numerous studies have confirmed that patients with IBD are at increased risk of VTE. 14 The absolute risk of VTE is highest during hospitalization where it is associated with 2-fold higher risk of mortality. 4,15 Thus, clinical practice guidelines have recommended pharmacological VTE prophylaxis for all hospitalized patients with IBD with the exception of those with hemodynamically significant gastrointestinal bleeding. 8 Prophylaxis rates among tertiary centers in our study far exceed those reported in a Canadian population-based study (including community hospitals) where VTE prophylaxis was,20%. 16 Our finding of differences in VTE prophylaxis among specialty services suggests there may be physician-based factors. For example, there may be greater VTE awareness among surgical services that has been previously described. 17 Variations in VTE prophylaxis across centers may reflect suboptimal quality of care, and may present an opportunity to introduce system-wide interventions, such as standardized IBD order sets, that may improve VTE prophylaxis. The use of electronic alert system and a pharmacist advocate were reported to be associated with increased implementation of VTE prophylaxis. 18,19 Interestingly, we found no difference in rates of VTE between those who did and did not receive VTE prophylaxis. Moreover, 79% of the mostly nonsurgical patients with VTE had received VTE prophylaxis, raising concern over its effectiveness. Although there have been previous reports of VTE occurring in 699

6 Nguyen et al Inflamm Bowel Dis Volume 23, Number 5, May 2017 hospitalized patients with IBD despite pharmacological prophylaxis, these have been in surgical patients, whereas most in our study who developed VTE on unfractionated or low molecular weight heparin were nonsurgical. While it is possible that these patients developed subclinical deep venous thrombosis before admission, a pilot study showed that the prevalence of asymptomatic deep venous thrombosis on admission to hospital is low. 20 However, just over a third of those who developed VTE who were administered prophylaxis received it within 24 hours of admission. In our study, those who received prophylaxis within 24 hours were half as likely to develop VTE as those who never received prophylaxis, although the event numbers were too small to achieve statistical significance. Our exploratory data, however, raises the potential importance of early timing of VTE prophylaxis, which needs to be confirmed in prospective studies. It was encouraging that gastrointestinal bleeding on admission was not a deterrent against early VTE prophylaxis. Interestingly, hemodynamically nonsignificant bleeding on admission was associated with higher implementation of early VTE prophylaxis, possibly because it is often accompanied by greater disease severity and perceived increase risk of VTE. Reassuringly, our study demonstrated that VTE prophylaxis did not increase the risk of major bleeding with or without transfusion. Testing for C. difficile also varied significantly across hospital sites and specialty. Surgeons, who achieved very high rates of VTE prophylaxis, ordered C. difficile testing only half as often as by gastroenterologists and general internists. Because C. difficile is usually medically treated, the deficit in testing may reflect relative lack of awareness among surgeons. The high prevalence of C. difficile positivity in our study population is concerning considering the rising trends in North America and the association of infection with higher mortality. 6 C. difficile testing is a cardinal step in the management of hospitalized UC because the early administration of antibiotics can contribute to response to medical therapy. 7 Although clinical guidelines recommend considering rescue medical therapy if there is inadequate response within 72 hours of steroid initiation, there is sparse published data documenting the promptness of administration of medical rescue therapy. 7 Because patients who have contraindications to or fail rescue therapy (e.g., infliximab or cyclosporine) will likely require colectomy, delays in initiative rescue therapy will prolong the inhospital time to surgery, which is associated with worse outcomes. 21 Prompt administration of medical rescue therapy is a complex process of care that involves multiple stages including prebiological work-up (e.g., tuberculosis testing), access to biological therapy, and patient consent. Thus, when measuring this indicator of care, there should be some consideration for leeway and defining timely administration as being within 7 days has been suggested. 11 Using a more stringent 4-days cutoff to define prompt treatment may be less feasible and unrealistic, as indicated by our data showing only 14% of admissions meeting this standard. Our study also showed geographic variation in bowel surgery during hospitalizations for an IBD flare, with the highest rates observed in Montreal. Regional variations in IBD-related surgery have been previously reported in the United States. 12,22 However, those studies were population-based studies that included both academic centers and community centers, large and small. Variations may have been multifactorial with contributions from relative access to medical and surgical interventions, patient and physician preferences, and case-mix. It was interesting to note that such variation exists even among large tertiary and quaternary IBD centers and may be a reflection of the same factors. One possible explanation is that there may be variations in clinical threshold for admission to hospital. For example, if Montreal had a higher threshold for hospitalization, then patients who are admitted may have relatively higher disease severity than other sites and may be at greater risk for surgery. An alternative hypothesis is that sites with higher surgery rates may have had a higher proportion or individuals who had previous exposure to one or more biologics. These individuals if unresponsive to steroids would be less likely to respond to second-line medical therapy and more likely to require surgery. Prospective studies are warranted to these hypotheses and further delineate the contributing factors to the variations in surgery rate. Among the limitations of our study was that it included only IBD referral centers from Canada s largest urban centers, limiting the generalizability of the findings. Because these centers have extensive experience treating high volumes of patients with IBD, the findings with respect to quality indicators may overestimate true rates at a population-based level. Another inherent limitation of retrospective studies is potential variation in documentation of indicators in the medical chart. For example, lack of steroid responsiveness data from 1 of the 4 sites limited our analysis for variations in timely administration of second-line therapy. Despite these limitations, we conducted a large multicenter study in major Canadian cities showing substantial geographic variations in quality indicators of care and outcomes. Although adherence to quality indicators were high overall, the hospitalbased variations may present an opportunity to improve quality of care through shared clinical care pathways that standardize processes of care. The nearly ubiquitous use of electronic order entry systems in hospitals may provide an opportunity to electronically integrate these pathways. The challenge remains developing a platform that is universally compatible with the diverse range of order entry systems used by hospitals throughout the country. The significant variations in quality of care among admitting services also highlights the role of education and awareness. One intriguing opportunity is empowering the patient through education to be a partner in assuring they receive high quality of care. Patients hospitalized for IBD represent a subgroup that has poor disease prognosis. Improving inpatient quality of care provides an important opportunity to optimize their health outcomes. REFERENCES 1. Longobardi T, Bernstein CN. Health care resource utilization in inflammatory bowel disease. Clin Gastroenterol Hepatol. 2006;4:

7 Inflamm Bowel Dis Volume 23, Number 5, May 2017 Inpatient IBD Quality of Care 2. Nguyen GC, Tuskey A, Dassopoulos T, et al. Rising hospitalization rates for inflammatory bowel disease in the United States between 1998 and Inflamm Bowel Dis. 2007;13: Longobardi T, Bernstein CN. Utilization of health-care resources by patients with IBD in Manitoba: a profile of time since diagnosis. Am J Gastroenterol. 2007;102: Nguyen GC, Sam J. Rising prevalence of venous thromboembolism and its impact on mortality among hospitalized inflammatory bowel disease patients. Am J Gastroenterol. 2008;103: Ananthakrishnan AN, McGinley EL, Binion DG. Excess hospitalisation burden associated with Clostridium difficile in patients with inflammatory bowel disease. Gut. 2008;57: Nguyen GC, Kaplan GG, Harris ML, et al. A national survey of the prevalence and impact of Clostridium difficile infection among hospitalized inflammatory bowel disease patients. Am J Gastroenterol. 2008;103: Bitton A, Buie D, Enns R, et al. Treatment of hospitalized adult patients with severe ulcerative colitis: Toronto consensus statements. Am J Gastroenterol. 2012;107: Nguyen GC, Bernstein CN, Bitton A, et al. Consensus statements on the risk, prevention, and treatment of venous thromboembolism in inflammatory bowel disease: Canadian Association of Gastroenterology. Gastroenterology. 2014; 146: Adult Inflammatory Bowel Disease Physician Performance Measures Set. Bethesda, MD: American Gastroenterological Association; 2011: Turner D, Walsh CM, Steinhart AH, et al. Response to corticosteroids in severe ulcerative colitis: a systematic review of the literature and a metaregression. Clin Gastroenterol Hepatol. 2007;5: Nguyen GC, Devlin SM, Afif W, et al. Defining quality indicators for best-practice management of inflammatory bowel disease in Canada. Can J Gastroenterol Hepatol. 2014;28: Nguyen GC, Laveist TA, Gearhart S, et al. Racial and geographic variations in colectomy rates among hospitalized ulcerative colitis patients. Clin Gastroenterol Hepatol. 2006;4: Statistics Canada: population and dwelling counts, for census metropolitan areas and census agglomerations, 2011 and 2006 censuses. Available at: Accessed October 28, Yuhara H, Steinmaus C, Corley D, et al. Meta-analysis: the risk of venous thromboembolism in patients with inflammatory bowel disease. Aliment Pharmacol Ther. 2013;37: Grainge MJ, West J, Card TR. Venous thromboembolism during active disease and remission in inflammatory bowel disease: a cohort study. Lancet. 2010;375: Kaplan GG, Lim A, Seow CH, et al. Colectomy is a risk factor for venous thromboembolism in ulcerative colitis. World J Gastroenterol. 2015;21: Tinsley A, Naymagon S, Enomoto LM, et al. Rates of pharmacologic venous thromboembolism prophylaxis in hospitalized patients with active ulcerative colitis: results from a tertiary care center. J Crohns Colitis. 2013;7:e635 e Mathers B, Williams E, Bedi G, et al. An electronic alert system is associated with a significant increase in pharmacologic venous thromboembolism prophylaxis rates among hospitalized inflammatory bowel disease patients. J Healthc Qual. [published online ahead of print May 5, 2016]. doi: /JHQ Ra G, Thanabalan R, Ratneswaran S, et al. Predictors and safety of venous thromboembolism prophylaxis among hospitalized inflammatory bowel disease patients. J Crohns Colitis. 2013;7:e479 e Nguyen GC, Wu H, Gulamhusein A, et al. The utility of screening for asymptomatic lower extremity deep venous thrombosis during inflammatory bowel disease flares: a pilot study. Inflamm Bowel Dis. 2013;19: Kaplan GG, McCarthy EP, Ayanian JZ, et al. Impact of hospital volume on postoperative morbidity and mortality following a colectomy for ulcerative colitis. Gastroenterology. 2008;134: Nguyen GC, Bayless TM, Powe NR, et al. Race and health insurance are predictors of hospitalized Crohn s disease patients undergoing bowel resection. Inflamm Bowel Dis. 2007;13: